[sdiy] VC Glide

[jhaible at t-online.de]

>     I didn't know I schriebed.... :-)

The wonders of German declension (;->)

> However, my 
> experience over the last 20 years with this part has been 
> that I don't 
> need to trim the devices to get good CV rejection.  I 
> remember the first 
> time I built something with one, a State Variable filter, 
> I was totally 
> blown away by the fact that the was not feed thru.  And 
> trimming the 
> OTA's in an SVF I always considered quite a pain, because 
> the trims 
> would interact if you didn't do it correctly.

This sounds convincing.

I wonder if it has anything to do with the "real" current
sources. On OTAs with resistor dividers, the input offset
voltage is multiplied to the output to the same degree
as the input signal is divided down (asuming unity gain
overall). So 1mV on the input of a 13600 can easily
become 20mV at the output (with diodes) and 1mV on
a 3080 (or 13600 without diodes) can be as much as
5% of the signal amplitude.

I have to look at the 3280 again. Maybe the "floating"
input scheme avoids this offset multiplying effect ...

JH.